From: Eric Niebler (eric_at_[hidden])
Date: 2004-01-05 12:33:06
Dave Harris wrote:
> In my own Windows code I avoid the names min and max altogether. Would it
> be feasible to add a boost::Min and boost::Max ? That is, are there any
> situations like the struct foo where max is part of the published boost
> Are there any other "known ways" which should be considered before
> updating the boost source?
Well, at least one compiler for Win32 (VC++) supports the push_macro and
pop_macro pragmas, but I don't know about the others. That's one
Another possible work-around would be to assign std::min/max to a
function pointer before invoking it. But that would be uglier and
probably have more portability issues that simply wrapping the
identifier in parens. And it doesn't help when min/max are member
functions in a public interface.
A third possibility is to simply roll with it:
max(a,b); // could be the macro or std::max!
That would be the Taoist's approach. :-)
I can't think of any others. boost::Min/Max would be a possibility, but
personally I'd prefer to use std::min/max if we can at all help it.
-- Eric Niebler Boost Consulting www.boost-consulting.com
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk